The harddrive went bad, and I dropped in a new one. Unfortunately the "recovery cd" wasn't readable!
My wife needed to do report cards (middleschool teach). The application used by the board is "Filemaker Pro", which only comes in Windows and Mac incarnations.
I put Redhat 9 on the laptop, and tested the report card application under WINE. Worked reasonably.
My wife got through the report cards, but I left Redhat on her system. She really doesn't notice -- it is more stable than Windows 98. Mozilla comes up slower than IE, and Evolution launches slower than Outlook Express, but she tolerates that.
Recently, she brought her laptop in to the school (report cards again). A co-worker remarked "What is that red hat in the corner?". Oh, my wife exclaimed, I guess this is -- Redhat Linux I'm using! Took months for her to realize she was using Linux. Others (Windows users) also don't really notice the difference.
The only complaints that I have had are:
1 - applications do load more slowly (this on machines with 128MB of RAM, CPU between 200 and 400Mhz, compared with W98).
2 - Redhat 9 spews a LOT of information when starting up, and (my wife finds this particularly amusing) gets very exited on shutdown (Halting! Stopping! Syncing! Flushing!...)
3 - Sometimes an application will "trip over itself". Eg. Mozilla won't exit completely. When this happens, Mozilla won't start again! (at least, not in normal time). Mozilla must be killed, and then restarted. Very confusing to my wife (although, turning the computer off and then on does restore things).
So, I agree that most users won't ever know the difference. Startup and shutdown could be quiet (and may be -- Redhat 9 is now "obsolete" -- so other distributions may have this better). Except for that, its much the same computer experience as any (recent 98 - present) Windows OS.
Theft of service? I am not being supplied a service! The RF is being fed to me whether or not I like it! Now, I don't decode it, because I couldn't be bothered.
But, if decoding the signal is "illegal" is it also illegal to measure that signal?
Patently not. If you or anyone else doesn't want me to have the signal, DON'T BEAM IT AT ME.
Got that? If its in MY FUCKING HOUSE, YOU GAVE IT TO ME.
Now, I *do* pay for cable -- go figure. But a BROADCAST SIGNAL?
If it where sufficiently strong, and I rectified it to power my stuff, would this be wrong? NO.
The sunshine that falls on my property is MINE. and the EM that crosses my property is also MINE.
Theft of "service". What a crock of shit that law is.
An interesting point. But, I bought 2 365's and a vrm from ebay. 200 bucks, a couple years back. Replaced my "dead" server - a p133. Couldn't find a replacement PS for it...
I kind of LIKE the "I need > 1ghz to surf the web" croud. Their cast-offs become my tools.
I browse the web, do email, a little image manipulation, and run a web server on my box.
I also run NIS, NFS, DHCP, Squid Proxy. I also run mailing lists, tape backup, and a cd burner.
I also run ftp, pop3 and smtp for a lan. Several times a day, the box fetches mail from several hotmail accounts, and alternate POP3. It also fetches and filters data from NNTP. It is also the NTP server for the LAN...
The box? a dual processor PPRO. 200Mhz with 128MB of RAM.
Works fine.
Client side? A 128MB PII 400. Works fine. Maybe one day I'll upgrade, but no reason to now.
An interesting test. I had noticed that KDE and GNOME terminal programs were "heavy-weight". I re-ran this on a PII 400 with 128MB RAM (Compaq Deskpro) running Redhat 9. I *also* added xterm to the mix:
[fred@raven fred]$ time for i in 0 1 2 3 4 5 6 7 8 9 ; do konsole -e date; done
real 0m38.170s user 0m17.140s sys 0m1.520s
[fred@raven fred]$ time for i in 0 1 2 3 4 5 6 7 8 9 ; do gnome-terminal -e date; done
real 0m20.002s user 0m8.240s sys 0m1.000s
[fred@raven fred]$ time for i in 0 1 2 3 4 5 6 7 8 9 ; do xterm -e date; done
real 0m9.578s user 0m4.110s sys 0m0.450s
gnome-terminal seems to start up twice as fast as konsole, and xterm twice as fast again.
Also, when scrolling, etc. xterm appears *much* lighter as well.
Now - I have 3 computers (yes, I *am* a developer). I do not buy at the "bleeding edge" -- all of my stuff is a few years old:
IBM Thinkpad 630E - 128MB Compaq Deskpro PII 400 - 128MB IBM PC 365 - dual PPRO 200 - 128MB
I firmly believe that these machines should be completely adequate to run an advanced desktop (except the 365 -- thats a server).
Upgrading these machines to more RAM *will* be painful -- but I don't want to discard them. One of the reasons that I am still using Redhat 9.
But even Redhat 9 is too bloated: After 1 day, the deskpro has grown its swapfile usage to around 100MB! This running Evolution, Mozilla, Xmms. and OpenOffice.
So, I guess the "sweet spot" is 256MB, even for Redhat 9. I find this a bit depressing, given that Mandrake 7 was able to play (with StarOffice, email, typical mix of applications) easily in 64MB.
Yes, I find additional features nice, but at the expense of 2 - 4 times the memory? Redhat 9 is not 2 to 4 times more capable than Mandrake 7. And Fedora Core 2 is not 2 times more capable than Redhat 9. So, as long as support is available for Redhat 9, I am not moving.
Ratboy.
NCR Invented SCSI?
on
The 3Com Saga
·
· Score: 2, Informative
SCSI was just a rework of SASI (and, in the very early days, they were compatible). And SASI was...
(drum role)
Shugart Associates Standard Interface
And, at the time, Shugart didn't have shit to do with NCR. Al Shugart started with IBM, and founded Shugart in 1973. He founded Seagate in '79.
Shugart teamed up with NCR in, what, 1981 to have ANSI standardized the interface, renaming it to SCSI.
But the "invention" belongs to Shugart, and not NCR.
In Unix (Linux) there is a difference between a MTA (Mail Transport Agent) and a MUA (Mail User Agent).
The idea is that the MUA can deal with many sources of mail, *without worrying about what they are*.
The "standard" MTA is sendmail (no flaming, please). Because of the divergence between "always on" peer to peer connections and "hosted" connections, MUAs have grown the ability to handle/some/ different kinds of mail (some MTA functionality). If you have an "always on" connection, it is possible to do *all* mail via the Unix "mail" (mailx) utility as the MUA.
Now, evolution is a "typical" MUA in that it support Unix mail/and/ POP3, etc. The POP3 support is for the hosted connection, where your system does NOT run an MTA at all.
But it doesn't have to go "all the way". Being a Linux (Unix) app, it is easy enough to run a normal MTA. Once you do this, it is very easy to integrate things like "HotMail" into evolution (and other POP3 mail, etc.)
For HotMail, aquire a script named "gotmail". Pure Perl, easy to set up. It can be run with cron, or manually. It will suck mail from HotMail (using that "horrid" web interface -- it uses the command line tool "curl") and send it to your MTA. From there, evolution reads the mail, and does its job.
Works fine for me. Indeed, I/also/ run a pop3 daemon, and evolution could access the Hotmail mail via POP3. I run POP3 because I have some clients that are Windows, and about the only "universal" model on Windows is the hosted model that relies on POP3.
My wish is that things like POP3 support would be OPTIONAL in evolution (making it faster to load). Local delivery would be fine. Still, the majority of users would have a need for POP3. They don't for Hotmail, but it is easy enough (the MUA/MTA division makes it trivial).
If you need Linux/Unix setup service, contact me at fred_weigel (at) hotmail.com My rates are quite reasonable.
Ratboy Looking for a driver dev in the Toronto Area? I'm available fred_weigel (at) hotmail.com
I use evolution for my email/calendaring needs. Works fine. Similar to Outlook. Needs a "connector" to interface with Exchange Server (which I haven't tried).
Swap partition should be "clean". There is no filesystem, and the kernel can (should) be allowed to assume clean i/o to swap (it is the same as memory).
At 128MB, swap is needed when loading OpenOffice (256MB would be fine). Not all used, and it runs ok with 128MB, but swap needed.
If there is a bad section on the swap partition, the kernel may be hitting it when starting OpenOffice. This causes a panic, and your system is (most likely) set to reboot when this happens (keeps 24/7 servers alive over such hardware failures).
So, bad swap => OpenOffice not loading, and a kernel panic. Kernel panic => automatic reboot.
All the best in getting this resolved -- if you want more help, email me: fred_weigel (at) hotmail.com will do (I check it daily).
Its not a clock multiplier. The idea is to use a common subfrequency of the two clocks: eg. FSB may be 100Mhz, Processor may be 400Mhz, a common subfrequency would be 4Mhz. This common subfrequency used to drive a PLL which can synchronize the clocks to reduce latencies.
Fred Langa is "right on" when comparing Microsoft OS offering against the attempted commercialized Linux offerings. From the perspective of a consumer.
He had the expectation that "all would work", and be "supported" (um... work) at least as well as Windows; given the the price was comparable.
From another perspective, that's wrong. If it doesn't work, get your money back -- that's what he paid for. But, Linux is a hobby system. If (or when) it works for Mr. Langa, he will know it, and use it.
What ticks me off is that Mr. Langa is being critical of Linux! You know, that hobby project. Get pissed at Suse, Redhat, Mandrake, (___ fill in the blank). Leave the hobbyist alone! Linux, Debian, et al. I am sure that Mr. Langa (as most of us) hs two standards -- one for professional atheletes, and another for amateur (Olympians, etc.). Yes, the professional bar is higher, as it should be.
If the F/OSS stuff is good enough, it will be used. Sure, criticize, but also give that community positive feedback. We aren't in it for money -- so some positive feedback would be useful.
The vilest thing that has happened to me in the Free Software world was a program I wrote (EMUL87). Distributed on SIMTEL; thousands of users. Not a word of positive feedback. Until one day (actually, 5 years later), when one consultant mailed me, and DEMAANDED I fix the software (because his client needed it). And if I didn't fix it IMMEDIATELY, I would be SUED. I told him to 'f off.
That nearly ended my relationship with F/OSS. But, I changed my mind. I like sharing, you see, and I get stuff from the community.
So, I feel that the F/OSS community is maligned and demotivated by the constant comparision with commercial software. The journalistic tack should be to take the commercial vendors to task if their offerings are so weak that F/OSS is actually competitive.
I understand why some people got defensive. Mr. Langa should CLEARLY state that the comparision is *not* with Linux or F/OSS, but with particular distributions or support organizations.
The Linux kernel runs on x86, mips, ibm mainframes, alpha, sparc and etc.
There are multiple versions of the kernel, smp and non-smp.
Typically, to load a driver module requires building it against the source tree for your particular kernel, which requires a kernel C compiler, and build environment to be installed.
A possible solution is to provide an architecture neutral layer (forth, byte-code) in which drivers could be written. A layer can introduce this to the kernel, giving only a single module to move. Perhaps existing drivers could be "compiled" into this intermediate (C -> driver language or byte-code).
Then, only the interface layer would need to be added into a Linux system -- and if built once, could be distributed as a binary. The "binary" drivers could be compiled into native code on the fly. Perhaps even x86 code could be used as the intermediate system (use something like QEMU to do the translation). Self-modifying code wouldn't be a problem. The "shim" layer for x86->x86 is simply an interception layer.
In principle, this could be done. And yes, I would use such a thing. As it is, whenever I upgrade a kernel, I have to rebuild all non-vendor modules (only 2 -- shfs and wlan-ng) and it is a pain in the ass.
Would sandboxing be of interest? Easier if not, but could be done if something like forth was used (declare the ports that the driver will touch -- provided for inspection to the operator, and trap all other references -- can also be filtered by kernel call).
Of course the shim would have to be GPLd, but would make proprietary drivers even easier than Windows (because they can be cross-platform).
Yes, this is a good idea.
Ratboy
PS I would be interested in working on this - contact me at fred (at) hotmail.com
Re:Not solution to slashdot effect, but still grea
on
Freecache
·
· Score: 2, Insightful
Actually, index.html would only be cached if it is 5MB or greater in size.
Which is unlikely. So it won't be cached. Nor will the PNG/GIFs.
We don't *care* about "proprietary, closed-source models". Because we don't do it for money, it's a hobby. Sure, Redhat, et.al. can make money, and that's good, but we would be doing it *anyway*.
What you should be asking is why the "proprietary, closed-source" (corporate model?) is at all threatened by an amateur job like Linux.
If the "reality" is that Microsoft (et.al.) are threatened by Linux, they have some very serious problems.
As I have previously stated -- when MS Longhorn is released I will look at it, and if it is better for my PROFESSIONAL needs, I will use it. If not, I will stick to Linux, (or whatever system meets those needs best).
If Microsoft can't win me over, it's actually Microsoft's problem. Nothing to do with Linux. And, seeing as how I *am* a computer hobbyist, I will contribute to the F/OSS. But, folks!, its just a hobby for me (and many, many others).
If the "hobby model" is so successful, maybe the IT area was TREMENDOUSLY overvalued. I'll even accept Ken's argument (silly on another level, but I will take it) that 85% of assests are now intangible "IP". The point is that the IP isn't worth anything! Sure, those companies are now very screwed. Maybe they will have to focus on MAKING PRODUCTS.
Calendars, schedules, meetings... yes, open source has 'em all. Predates MS Exchange by a considerable degree.
Let me wind back... "finger user". Ah well, Exchange won't finger people to get plans. Calender and reminder services for Unix date way, WAY back. Just that Exchange doesn't support them.
So "we" (Linux/Unix/other OSs) have to kludge in support for MS.
I also use Linux "full-time". Except for Exchange. I have a machine on my desktop that is DEDICATED to Exchange Client. Nothing else. Of course, "Outlook" has been paid for, so that was it. Until now. I had the extra machine, a KVM, and even a network hub pretty much devoted to the "Exchange Experience".
All because Exchange can't be accessed via HTTP internally (here, at any rate), and won't "finger" me.
So, this is great news. I can finally get back a vast chunk of my desktop. But I still wish that the damn Exchange server would do more "compatible" (and, for me, useful) stuff.
You see, with biometrics, a compromise is FOREVER.
You are issued a card, with your biometrics encoded on it. Assume it is not tied to a central authority. The card can be "forged", and the biometrics add nothing to the security. We are talking about "illegal immigration" here, which a person will pay many thousands of dollars for. If it costs in excess of $1000 to forge the local biometric card, and the card can be issued for sufficiently low funds (say, under $20), then I would contemplate the measure.
Assume that the biometrics are tied to a central authority (and this is what is being proposed). If the data held by the central authority can be altered, people can be screwed. For life. Your identity could be co-opted by someone inserting other biometric data, or your biometric data could be associated with someone else.
Now, you are another person, with no real way of repudiating. This is not a password that can be changed, or a PIN number. It is *your* biometric data. If, suddenly, you (your biometrics) are on the FBI most-wanted... well... I'll leave it to you.
The question is then: do you trust the security of the central authority(ies)? The "UK" database, the "US" database, and "CDN" database? Do you believe that social engineering attacks won't be mounted or will fail? Because once the alteration is done, it will be almost impossible to undo.
Especially if the biometric id is taken to be good enough to be universal. The more it is relied on, the more at risk the central authority is.
And as your "n% becomes smaller and smaller" the risk approaches 100% (from my perspective -- you may be a genius security professional who can sucessfully 'vette such a system -- I am just a code monkey who has dealt with security).
1 - avoid lock-in to one particular processor. I have a PPRO dual processor rig, Pentium II, Thinkpad and Pentium 1. Does the same optimization work on all of these? [Answer: NO]. I have also had a quad Sparc, and MIPS based machines (all with Linux).
2 - profiling and JIT can result in better code than static compilation and optimization. The optimization *should* be directed toward the applictions use. Something that can be done by running the app, and then rebuilding the app. Or, use a JIT (ref. HP work - binding an emulator and JIT optimizer to PA-RISC *unoptimized* binaries gives results comparable to fully optimized binaries. At a fraction of the compile time.)
3 - Why would an app delivery platform need all those compilers anyway?
Next, use of mono. You think its bad that mono is actually *used*?
We are talking about Gnome using mono, not mono using Gnome. Very easy, anyway. Nothing prevents you from coding pieces of Gnome in *any* language (as long as you've got CORBA).
Anyway, mono may be an application delivery mechanism for the web, which would imply the possibility of "viruses". Which needs to be taken care of. Sandboxing is probably the only viable mechanism. Games? If its a RAD system, most programmers will write a utility or game just to "get their feet wet". Nothing wrong with that.
And now, on to your objection about Win32 ports.
The source is open. If someone wants it, they can do it. F/OSS is *not* a monolithic thing. You want to restrict the use of software based on.. what.. your perception that Windows is in some way a lesser platform, and you want to force it, by not making some software available, even though the platform is capable? Shame on you.
"At present we dont have ways in place for people to watch digital movies and othe rprotected content in ways the the owners are willing to produce or share thier content for. Let's not get into an RIAA riff here. The point is that lots of people do want to "rent" content and watch it and without a secure communication channel they cant."
I call bullshit. The owners WILL produce and share content. They would like more control, but will be willing to settle for what they have. Just stop buying (licensing?) or renting content for a while and see what happens.
We *have* secure communications channels. We have available good encryption. What the cartels want is CONTROL. Because they don't see the need to trust their customers. They could sell me a movie or song over the 'net, securely and conveniently, but they won't -- because they don't trust me.
So fuck them -- just ignore them. They'll come around eventually. Unless this "Palladium" crap wins.
*How* can NGSCB and Palladium be used to enforce the GPL?
Oh, by tying the source code to a key, which makes it impossible to change the source code and use the same key... but the verification is against the key. By tying the binary to a key, and making it impossible to modify the binary? So, rebuild the binary, and key use is lost.
In other words, these measures *can't* be used to enforce GPL. So much for this tool.
Now, is Palladium a security project? Well, yes, but not for the end user. Indeed, the end user can run the same old trojans, etc. as before. Palladium *will* prevent the trojan from accessing data that has bee "protected", by kicking out the unsuitable software.
It was NEVER meant to secure YOUR stuff -- if you want that, go use GPG, etc. I assume that even MS Outlook must have some integration with GPG! (all of my emails are digitally signed).
As I have mentioned before, I use a hotmail account. Any spam uses MSN bandwidth, and not my bandwidth. I set the automatic spam filter to DELETE.
If you and I establish a relationship, you get my real email. And, yes, I do have my own domain(s). Indeed, you could probably guess my real email.
I do this to keep my spam bandwidth low. I have tried blacklisting, whitelisting, RBLs, but they don't work. Hotmail takes care of most of the spam problem, and not on my dime. My ISP doesn't like servers, but turns a blind eye to low volume users (which I try to be).
And now, an UNPUBLISHED email address on my domain is being spammed. I believe that the address was in someones address book, and got scraped. After a single week, spam to email ratio has hit 9:1 (and getting worse).
I could disable the email, but will get hits for a long time anyway... maybe I'll just disable the damn domain for a few months...
People are welcome to information from my site. On average, 300MB of month of programming manuals and documentation served. But the fucking spammers. Spoil it for everyone.
So, I consider it very professional (and you can always send email to my domain using the common addresses for administration -- interesting that those DON'T get spammed!)
Yes, Windows Server 2003 will do it. The new Terminal Services can redirect audio.
Of course, Windows Server 2003 is US $650 (give or take). It wasn't clear (as the audio redirection is a new feature introduced with Windows Server 2003) that it would work with other versions.
The parent is REALLY good troll. Read it and learn.
And, I'm going to fall for it...
ESD - 1 second lag with MP3s. Say what?!? Not true. Yes, its a "user" daemon, and that is as it should be. Because it also supports network transparency for audio. Eg. Try playing audio on your laptop... with the sound coming out of your stereo... that happens to be wired to your media server... Come on, I dare you, do it in Windows.
Now, lets have the Linux answer (drum roll)
export ESPEAKER=media.lan:16001
The Windows solutions are just SO LAME.
Video cards? How about any fucking thing with VESA support. Good enough for you? And that's what Dell, HP, et al sell. No, the corporate desktop doesn't have need for the latest and greatest gaming card. This is FUD at its finest. Back up your claim, FUD-boy. Name a system *name-brand* with generic video that Linux (redhat 9, or ES) does not support suitably for corporate use.
And last, Linux is *not* a product. It's a kernel. Stick with PRODUCT vendors, and your last gastly argument goes away... Its the same as saying "Oh dear, my Windows box won't run my cell-phone software". Each of your "100's of distributions" is a "competing yet compatible" product.
So, my final retort is that you are the most excellent FUD-boy troller I have seen this week! I am in awe of your mad skillz, sensei!
Please put paragraph breaks in. Makes it easier to follow what you are trying to say...
"Why shouldn't these artists get paid for the time and effort? We pay bankers to handle our money, cooks to make our dinner, maids to clean our houses, but we can't pay artists who actually make our lives enjoyable?"
I do go to concerts. I spend about $1000 per year on: concerts, musical theatre, movies, and sports. That is more than I pay to maids (4 visits per year, $100 per visit).
When I do something, I don't get paid after it is finished. Doesn't matter how many people use it. Yes, you my have my code in your computer RIGHT NOW. But I don't get royalties. And that's ok with me... Now, the musicians do get royalties. I would think that 1 cent over a million uses is still 100,000 dollars. More than I make, anyway. And for this, do nothing.
As to THIS being the "system of patronage extends back for countless millennia"?
I don't think so. 100 years ago the only way to hear an artist was to be there. Or listen to someone else play the music. Works *were* commissioned. Certainly, no one made money on selling recordings -- there weren't any.
I don't mind people making money selling recordings... just get it into line with costs and reasonable profit. Or I won't buy. $1 per song is FAR too much. Given that 10 cents a song pays for royalties, encoding, distribution *and* profit.
If you insist on pricing electronic recordings at $1 a song, and CDs at $15, I won't buy. [Note - WalMart - usually not known for generosity - sells some DVD movies at $4.88 CDN].
I will still go to concerts. Am I selfish and cheap? Thanks for thinking that about me, but I have purchased more CDs and sound recodings than the entire human race did before 1850. More, even. I think that make me really generous to the artists.
Slashdot Mirror
My wife uses an IBM Thinkpad (360E).
The harddrive went bad, and I dropped in a new one. Unfortunately the "recovery cd" wasn't readable!
My wife needed to do report cards (middleschool teach). The application used by the board is "Filemaker Pro", which only comes in Windows and Mac incarnations.
I put Redhat 9 on the laptop, and tested the report card application under WINE. Worked reasonably.
My wife got through the report cards, but I left Redhat on her system. She really doesn't notice -- it is more stable than Windows 98. Mozilla comes up slower than IE, and Evolution launches slower than Outlook Express, but she tolerates that.
Recently, she brought her laptop in to the school (report cards again). A co-worker remarked "What is that red hat in the corner?". Oh, my wife exclaimed, I guess this is -- Redhat Linux I'm using! Took months for her to realize she was using Linux. Others (Windows users) also don't really notice the difference.
The only complaints that I have had are:
1 - applications do load more slowly (this on machines with 128MB of RAM, CPU between 200 and 400Mhz, compared with W98).
2 - Redhat 9 spews a LOT of information when starting up, and (my wife finds this particularly amusing) gets very exited on shutdown (Halting! Stopping! Syncing! Flushing!...)
3 - Sometimes an application will "trip over itself". Eg. Mozilla won't exit completely. When this happens, Mozilla won't start again! (at least, not in normal time). Mozilla must be killed, and then restarted. Very confusing to my wife (although, turning the computer off and then on does restore things).
So, I agree that most users won't ever know the difference. Startup and shutdown could be quiet (and may be -- Redhat 9 is now "obsolete" -- so other distributions may have this better). Except for that, its much the same computer experience as any (recent 98 - present) Windows OS.
Ratboy.
Theft of service? I am not being supplied a service! The RF is being fed to me whether or not I like it! Now, I don't decode it, because I couldn't be bothered.
But, if decoding the signal is "illegal" is it also illegal to measure that signal?
Patently not. If you or anyone else doesn't want me to have the signal, DON'T BEAM IT AT ME.
Got that? If its in MY FUCKING HOUSE, YOU GAVE IT TO ME.
Now, I *do* pay for cable -- go figure. But a BROADCAST SIGNAL?
If it where sufficiently strong, and I rectified it to power my stuff, would this be wrong? NO.
The sunshine that falls on my property is MINE. and the EM that crosses my property is also MINE.
Theft of "service". What a crock of shit that law is.
Ratboy.
An interesting point. But, I bought 2 365's and a vrm from ebay. 200 bucks, a couple years back. Replaced my "dead" server - a p133. Couldn't find a replacement PS for it...
I kind of LIKE the "I need > 1ghz to surf the web" croud. Their cast-offs become my tools.
ratboy
I browse the web, do email, a little image manipulation, and run a web server on my box.
I also run NIS, NFS, DHCP, Squid Proxy. I also run mailing lists, tape backup, and a cd burner.
I also run ftp, pop3 and smtp for a lan. Several times a day, the box fetches mail from several hotmail accounts, and alternate POP3. It also fetches and filters data from NNTP. It is also the NTP server for the LAN...
The box? a dual processor PPRO. 200Mhz with 128MB of RAM.
Works fine.
Client side? A 128MB PII 400. Works fine. Maybe one day I'll upgrade, but no reason to now.
Ratboy.
An interesting test. I had noticed that KDE and GNOME terminal programs were "heavy-weight". I re-ran this on a PII 400 with 128MB RAM (Compaq Deskpro) running Redhat 9. I *also* added xterm to the mix:
[fred@raven fred]$ time for i in 0 1 2 3 4 5 6 7 8 9 ; do konsole -e date; done
real 0m38.170s
user 0m17.140s
sys 0m1.520s
[fred@raven fred]$ time for i in 0 1 2 3 4 5 6 7 8 9 ; do gnome-terminal -e date; done
real 0m20.002s
user 0m8.240s
sys 0m1.000s
[fred@raven fred]$ time for i in 0 1 2 3 4 5 6 7 8 9 ; do xterm -e date; done
real 0m9.578s
user 0m4.110s
sys 0m0.450s
gnome-terminal seems to start up twice as fast as konsole, and xterm twice as fast again.
Also, when scrolling, etc. xterm appears *much* lighter as well.
Now - I have 3 computers (yes, I *am* a developer). I do not buy at the "bleeding edge" -- all of my stuff is a few years old:
IBM Thinkpad 630E - 128MB
Compaq Deskpro PII 400 - 128MB
IBM PC 365 - dual PPRO 200 - 128MB
I firmly believe that these machines should be completely adequate to run an advanced desktop (except the 365 -- thats a server).
Upgrading these machines to more RAM *will* be painful -- but I don't want to discard them. One of the reasons that I am still using Redhat 9.
But even Redhat 9 is too bloated: After 1 day, the deskpro has grown its swapfile usage to around 100MB! This running Evolution, Mozilla, Xmms. and OpenOffice.
So, I guess the "sweet spot" is 256MB, even for Redhat 9. I find this a bit depressing, given that Mandrake 7 was able to play (with StarOffice, email, typical mix of applications) easily in 64MB.
Yes, I find additional features nice, but at the expense of 2 - 4 times the memory? Redhat 9 is not 2 to 4 times more capable than Mandrake 7. And Fedora Core 2 is not 2 times more capable than Redhat 9. So, as long as support is available for Redhat 9, I am not moving.
Ratboy.
SCSI was just a rework of SASI (and, in the very early days, they were compatible). And SASI was...
(drum role)
Shugart Associates Standard Interface
And, at the time, Shugart didn't have shit to do with NCR. Al Shugart started with IBM, and founded Shugart in 1973. He founded Seagate in '79.
Shugart teamed up with NCR in, what, 1981 to have ANSI standardized the interface, renaming it to SCSI.
But the "invention" belongs to Shugart, and not NCR.
Ratboy.
Where do I start?
/some/ different kinds of mail (some MTA functionality). If you have an "always on" connection, it is possible to do *all* mail via the Unix "mail" (mailx) utility as the MUA.
/and/ POP3, etc. The POP3 support is for the hosted connection, where your system does NOT run an MTA at all.
/also/ run a pop3 daemon, and evolution could access the Hotmail mail via POP3. I run POP3 because I have some clients that are Windows, and about the only "universal" model on Windows is the hosted model that relies on POP3.
At the beginning, I suppose...
In Unix (Linux) there is a difference between a MTA (Mail Transport Agent) and a MUA (Mail User Agent).
The idea is that the MUA can deal with many sources of mail, *without worrying about what they are*.
The "standard" MTA is sendmail (no flaming, please). Because of the divergence between "always on" peer to peer connections and "hosted" connections, MUAs have grown the ability to handle
Now, evolution is a "typical" MUA in that it support Unix mail
But it doesn't have to go "all the way". Being a Linux (Unix) app, it is easy enough to run a normal MTA. Once you do this, it is very easy to integrate things like "HotMail" into evolution (and other POP3 mail, etc.)
For HotMail, aquire a script named "gotmail". Pure Perl, easy to set up. It can be run with cron, or manually. It will suck mail from HotMail (using that "horrid" web interface -- it uses the command line tool "curl") and send it to your MTA. From there, evolution reads the mail, and does its job.
Works fine for me. Indeed, I
My wish is that things like POP3 support would be OPTIONAL in evolution (making it faster to load). Local delivery would be fine. Still, the majority of users would have a need for POP3. They don't for Hotmail, but it is easy enough (the MUA/MTA division makes it trivial).
If you need Linux/Unix setup service, contact me at fred_weigel (at) hotmail.com My rates are quite reasonable.
Ratboy
Looking for a driver dev in the Toronto Area? I'm available fred_weigel (at) hotmail.com
Not an "Outlook Clone", but it works --
I use evolution for my email/calendaring needs. Works fine. Similar to Outlook. Needs a "connector" to interface with Exchange Server (which I haven't tried).
Ratboy.
Swap partition should be "clean". There is no filesystem, and the kernel can (should) be allowed to assume clean i/o to swap (it is the same as memory).
At 128MB, swap is needed when loading OpenOffice (256MB would be fine). Not all used, and it runs ok with 128MB, but swap needed.
If there is a bad section on the swap partition, the kernel may be hitting it when starting OpenOffice. This causes a panic, and your system is (most likely) set to reboot when this happens (keeps 24/7 servers alive over such hardware failures).
So, bad swap => OpenOffice not loading, and a kernel panic. Kernel panic => automatic reboot.
All the best in getting this resolved -- if you want more help, email me: fred_weigel (at) hotmail.com will do (I check it daily).
Ratboy.
I'll fix it for you...
I assume that the box has 128MB of memory. If it has more, the issue is probably elsewhere.
I think there is a bad spot on the swapfile. Replace the hard drive. Also, boost memory to 256MB.
Should be fine then.
Ratboy.
ALL Computers Inc.
Mers Kutt
Taught Computer Science *before* there was such a thing -- Math professor at Queens.
Created the first "personal computer" -- the MCM/70 in 1971. Ran APL, so not very popular among the home set.
Created the first circuit to add memory management and virtual memory for 8086 based PCs.
Won "Upgrade Product of the Year" from Ziff-Davis for the ALL/Chargecard.
Made CPU upgrades for older PCs (286, 386 to 486).
Pretty well done now (I don't think Mers will actually "retire" though). If Intel infringed, I guess Mers will get his pay-out. Good for him.
Ratboy.
Its not a clock multiplier. The idea is to use a common subfrequency of the two clocks: eg. FSB may be 100Mhz, Processor may be 400Mhz, a common subfrequency would be 4Mhz. This common subfrequency used to drive a PLL which can synchronize the clocks to reduce latencies.
Ratboy.
Fred Langa is "right on" when comparing Microsoft OS offering against the attempted commercialized Linux offerings. From the perspective of a consumer.
He had the expectation that "all would work", and be "supported" (um... work) at least as well as Windows; given the the price was comparable.
From another perspective, that's wrong. If it doesn't work, get your money back -- that's what he paid for. But, Linux is a hobby system. If (or when) it works for Mr. Langa, he will know it, and use it.
What ticks me off is that Mr. Langa is being critical of Linux! You know, that hobby project. Get pissed at Suse, Redhat, Mandrake, (___ fill in the blank). Leave the hobbyist alone! Linux, Debian, et al. I am sure that Mr. Langa (as most of us) hs two standards -- one for professional atheletes, and another for amateur (Olympians, etc.). Yes, the professional bar is higher, as it should be.
If the F/OSS stuff is good enough, it will be used. Sure, criticize, but also give that community positive feedback. We aren't in it for money -- so some positive feedback would be useful.
The vilest thing that has happened to me in the Free Software world was a program I wrote (EMUL87). Distributed on SIMTEL; thousands of users. Not a word of positive feedback. Until one day (actually, 5 years later), when one consultant mailed me, and DEMAANDED I fix the software (because his client needed it). And if I didn't fix it IMMEDIATELY, I would be SUED. I told him to 'f off.
That nearly ended my relationship with F/OSS. But, I changed my mind. I like sharing, you see, and I get stuff from the community.
So, I feel that the F/OSS community is maligned and demotivated by the constant comparision with commercial software. The journalistic tack should be to take the commercial vendors to task if their offerings are so weak that F/OSS is actually competitive.
I understand why some people got defensive. Mr. Langa should CLEARLY state that the comparision is *not* with Linux or F/OSS, but with particular distributions or support organizations.
Enough of a rant.
Ratboy.
The Linux kernel runs on x86, mips, ibm mainframes, alpha, sparc and etc.
There are multiple versions of the kernel, smp and non-smp.
Typically, to load a driver module requires building it against the source tree for your particular kernel, which requires a kernel C compiler, and build environment to be installed.
A possible solution is to provide an architecture neutral layer (forth, byte-code) in which drivers could be written. A layer can introduce this to the kernel, giving only a single module to move. Perhaps existing drivers could be "compiled" into this intermediate (C -> driver language or byte-code).
Then, only the interface layer would need to be added into a Linux system -- and if built once, could be distributed as a binary. The "binary" drivers could be compiled into native code on the fly. Perhaps even x86 code could be used as the intermediate system (use something like QEMU to do the translation). Self-modifying code wouldn't be a problem. The "shim" layer for x86->x86 is simply an interception layer.
In principle, this could be done. And yes, I would use such a thing. As it is, whenever I upgrade a kernel, I have to rebuild all non-vendor modules (only 2 -- shfs and wlan-ng) and it is a pain in the ass.
Would sandboxing be of interest? Easier if not, but could be done if something like forth was used (declare the ports that the driver will touch -- provided for inspection to the operator, and trap all other references -- can also be filtered by kernel call).
Of course the shim would have to be GPLd, but would make proprietary drivers even easier than Windows (because they can be cross-platform).
Yes, this is a good idea.
Ratboy
PS I would be interested in working on this - contact me at fred (at) hotmail.com
Actually, index.html would only be cached if it is 5MB or greater in size.
Which is unlikely. So it won't be cached. Nor will the PNG/GIFs.
Ratboy
We don't *care* about "proprietary, closed-source models". Because we don't do it for money, it's a hobby. Sure, Redhat, et.al. can make money, and that's good, but we would be doing it *anyway*.
What you should be asking is why the "proprietary, closed-source" (corporate model?) is at all threatened by an amateur job like Linux.
If the "reality" is that Microsoft (et.al.) are threatened by Linux, they have some very serious problems.
As I have previously stated -- when MS Longhorn is released I will look at it, and if it is better for my PROFESSIONAL needs, I will use it. If not, I will stick to Linux, (or whatever system meets those needs best).
If Microsoft can't win me over, it's actually Microsoft's problem. Nothing to do with Linux. And, seeing as how I *am* a computer hobbyist, I will contribute to the F/OSS. But, folks!, its just a hobby for me (and many, many others).
If the "hobby model" is so successful, maybe the IT area was TREMENDOUSLY overvalued. I'll even accept Ken's argument (silly on another level, but I will take it) that 85% of assests are now intangible "IP". The point is that the IP isn't worth anything! Sure, those companies are now very screwed. Maybe they will have to focus on MAKING PRODUCTS.
Ratboy
????
In a word, WRONG.
Calendars, schedules, meetings... yes, open source has 'em all. Predates MS Exchange by a considerable degree.
Let me wind back... "finger user". Ah well, Exchange won't finger people to get plans. Calender and reminder services for Unix date way, WAY back. Just that Exchange doesn't support them.
So "we" (Linux/Unix/other OSs) have to kludge in support for MS.
I also use Linux "full-time". Except for Exchange. I have a machine on my desktop that is DEDICATED to Exchange Client. Nothing else. Of course, "Outlook" has been paid for, so that was it. Until now. I had the extra machine, a KVM, and even a network hub pretty much devoted to the "Exchange Experience".
All because Exchange can't be accessed via HTTP internally (here, at any rate), and won't "finger" me.
So, this is great news. I can finally get back a vast chunk of my desktop. But I still wish that the damn Exchange server would do more "compatible" (and, for me, useful) stuff.
YMMV
Ratboy.
You see, with biometrics, a compromise is FOREVER.
You are issued a card, with your biometrics encoded on it. Assume it is not tied to a central authority. The card can be "forged", and the biometrics add nothing to the security. We are talking about "illegal immigration" here, which a person will pay many thousands of dollars for. If it costs in excess of $1000 to forge the local biometric card, and the card can be issued for sufficiently low funds (say, under $20), then I would contemplate the measure.
Assume that the biometrics are tied to a central authority (and this is what is being proposed). If the data held by the central authority can be altered, people can be screwed. For life. Your identity could be co-opted by someone inserting other biometric data, or your biometric data could be associated with someone else.
Now, you are another person, with no real way of repudiating. This is not a password that can be changed, or a PIN number. It is *your* biometric data. If, suddenly, you (your biometrics) are on the FBI most-wanted... well... I'll leave it to you.
The question is then: do you trust the security of the central authority(ies)? The "UK" database, the "US" database, and "CDN" database? Do you believe that social engineering attacks won't be mounted or will fail? Because once the alteration is done, it will be almost impossible to undo.
Especially if the biometric id is taken to be good enough to be universal. The more it is relied on, the more at risk the central authority is.
And as your "n% becomes smaller and smaller" the risk approaches 100% (from my perspective -- you may be a genius security professional who can sucessfully 'vette such a system -- I am just a code monkey who has dealt with security).
Ratboy
Reasons for JIT:
1 - avoid lock-in to one particular processor. I have a PPRO dual processor rig, Pentium II, Thinkpad and Pentium 1. Does the same optimization work on all of these? [Answer: NO]. I have also had a quad Sparc, and MIPS based machines (all with Linux).
2 - profiling and JIT can result in better code than static compilation and optimization. The optimization *should* be directed toward the applictions use. Something that can be done by running the app, and then rebuilding the app. Or, use a JIT (ref. HP work - binding an emulator and JIT optimizer to PA-RISC *unoptimized* binaries gives results comparable to fully optimized binaries. At a fraction of the compile time.)
3 - Why would an app delivery platform need all those compilers anyway?
Next, use of mono. You think its bad that mono is actually *used*?
We are talking about Gnome using mono, not mono using Gnome. Very easy, anyway. Nothing prevents you from coding pieces of Gnome in *any* language (as long as you've got CORBA).
Anyway, mono may be an application delivery mechanism for the web, which would imply the possibility of "viruses". Which needs to be taken care of. Sandboxing is probably the only viable mechanism. Games? If its a RAD system, most programmers will write a utility or game just to "get their feet wet". Nothing wrong with that.
And now, on to your objection about Win32 ports.
The source is open. If someone wants it, they can do it. F/OSS is *not* a monolithic thing. You want to restrict the use of software based on.. what.. your perception that Windows is in some way a lesser platform, and you want to force it, by not making some software available, even though the platform is capable? Shame on you.
Ratboy
"At present we dont have ways in place for people to watch digital movies and othe rprotected content in ways the the owners are willing to produce or share thier content for. Let's not get into an RIAA riff here. The point is that lots of people do want to "rent" content and watch it and without a secure communication channel they cant."
I call bullshit. The owners WILL produce and share content. They would like more control, but will be willing to settle for what they have. Just stop buying (licensing?) or renting content for a while and see what happens.
We *have* secure communications channels. We have available good encryption. What the cartels want is CONTROL. Because they don't see the need to trust their customers. They could sell me a movie or song over the 'net, securely and conveniently, but they won't -- because they don't trust me.
So fuck them -- just ignore them. They'll come around eventually. Unless this "Palladium" crap wins.
Ratboy.
I call bullshit.
*How* can NGSCB and Palladium be used to enforce the GPL?
Oh, by tying the source code to a key, which makes it impossible to change the source code and use the same key... but the verification is against the key. By tying the binary to a key, and making it impossible to modify the binary? So, rebuild the binary, and key use is lost.
In other words, these measures *can't* be used to enforce GPL. So much for this tool.
Now, is Palladium a security project? Well, yes, but not for the end user. Indeed, the end user can run the same old trojans, etc. as before. Palladium *will* prevent the trojan from accessing data that has bee "protected", by kicking out the unsuitable software.
It was NEVER meant to secure YOUR stuff -- if you want that, go use GPG, etc. I assume that even MS Outlook must have some integration with GPG! (all of my emails are digitally signed).
Ratboy.
As I have mentioned before, I use a hotmail account. Any spam uses MSN bandwidth, and not my bandwidth. I set the automatic spam filter to DELETE.
If you and I establish a relationship, you get my real email. And, yes, I do have my own domain(s). Indeed, you could probably guess my real email.
I do this to keep my spam bandwidth low. I have tried blacklisting, whitelisting, RBLs, but they don't work. Hotmail takes care of most of the spam problem, and not on my dime. My ISP doesn't like servers, but turns a blind eye to low volume users (which I try to be).
And now, an UNPUBLISHED email address on my domain is being spammed. I believe that the address was in someones address book, and got scraped. After a single week, spam to email ratio has hit 9:1 (and getting worse).
I could disable the email, but will get hits for a long time anyway... maybe I'll just disable the damn domain for a few months...
People are welcome to information from my site. On average, 300MB of month of programming manuals and documentation served. But the fucking spammers. Spoil it for everyone.
So, I consider it very professional (and you can always send email to my domain using the common addresses for administration -- interesting that those DON'T get spammed!)
Ratboy
rkz
Yes, Windows Server 2003 will do it. The new Terminal Services can redirect audio.
Of course, Windows Server 2003 is US $650 (give or take). It wasn't clear (as the audio redirection is a new feature introduced with Windows Server 2003) that it would work with other versions.
How much did your software end up costing?
Ratboy.
The parent is REALLY good troll. Read it and learn.
And, I'm going to fall for it...
ESD - 1 second lag with MP3s. Say what?!? Not true. Yes, its a "user" daemon, and that is as it should be. Because it also supports network transparency for audio. Eg. Try playing audio on your laptop... with the sound coming out of your stereo... that happens to be wired to your media server... Come on, I dare you, do it in Windows.
Now, lets have the Linux answer (drum roll)
export ESPEAKER=media.lan:16001
The Windows solutions are just SO LAME.
Video cards? How about any fucking thing with VESA support. Good enough for you? And that's what Dell, HP, et al sell. No, the corporate desktop doesn't have need for the latest and greatest gaming card. This is FUD at its finest. Back up your claim, FUD-boy. Name a system *name-brand* with generic video that Linux (redhat 9, or ES) does not support suitably for corporate use.
And last, Linux is *not* a product. It's a kernel. Stick with PRODUCT vendors, and your last gastly argument goes away... Its the same as saying "Oh dear, my Windows box won't run my cell-phone software". Each of your "100's of distributions" is a "competing yet compatible" product.
So, my final retort is that you are the most excellent FUD-boy troller I have seen this week! I am in awe of your mad skillz, sensei!
Ratboy.
Please put paragraph breaks in. Makes it easier to follow what you are trying to say...
"Why shouldn't these artists get paid for the time and effort? We pay bankers to handle our money, cooks to make our dinner, maids to clean our houses, but we can't pay artists who actually make our lives enjoyable?"
I do go to concerts. I spend about $1000 per year on: concerts, musical theatre, movies, and sports. That is more than I pay to maids (4 visits per year, $100 per visit).
When I do something, I don't get paid after it is finished. Doesn't matter how many people use it. Yes, you my have my code in your computer RIGHT NOW. But I don't get royalties. And that's ok with me... Now, the musicians do get royalties. I would think that 1 cent over a million uses is still 100,000 dollars. More than I make, anyway. And for this, do nothing.
As to THIS being the "system of patronage extends back for countless millennia"?
I don't think so. 100 years ago the only way to hear an artist was to be there. Or listen to someone else play the music. Works *were* commissioned. Certainly, no one made money on selling recordings -- there weren't any.
I don't mind people making money selling recordings... just get it into line with costs and reasonable profit. Or I won't buy. $1 per song is FAR too much. Given that 10 cents a song pays for royalties, encoding, distribution *and* profit.
If you insist on pricing electronic recordings at $1 a song, and CDs at $15, I won't buy. [Note - WalMart - usually not known for generosity - sells some DVD movies at $4.88 CDN].
I will still go to concerts. Am I selfish and cheap? Thanks for thinking that about me, but I have purchased more CDs and sound recodings than the entire human race did before 1850. More, even. I think that make me really generous to the artists.
Ratboy.